Introduction

The Maxvision Image Retrieval Server delivers AI-powered image analysis services through two flexible modes: a visual retrieval interface and a developer-oriented algorithm API.

The visual interface provides an intuitive user experience for essential tasks such as object detection, feature extraction, and similarity matching.
The algorithm API offers a more comprehensive suite of capabilities—including object detection, attribute recognition, feature extraction, image indexing, as well as both 1:1 and 1:N comparison—ensuring high algorithmic accuracy while supporting large-scale concurrent requests.

Built on a microservices-based distributed architecture, the server is designed for both standalone and scalable deployments. It maximizes computing efficiency to deliver fast, accurate, and stable analysis of massive surveillance image streams—making it an ideal solution for complex environments such as industrial parks, retail hubs, airports, and border control facilities where intelligent visual recognition and matching are critical.

Features

Compatible with Mainstream Hardware and Software

Multiple AI Acceleration Options

Broad Model Compatibility

High-Performance, High-Accuracy Algorithms

Flexible Deployment Modes

Parameters

CPU

≥16 cores @ 1.9 GHz

AI Processing Power

≥ 176T (INT8)

eMMC

≥ 64GB

Memory

≥ 48GB

Enterprise-grade Storage

≥ 2TB

Interfaces

At least: 2 × 10G SFP+ fiber ports 4 × USB 3.0 2 × 1G RJ45 Ethernet ports 1 × RJ45 serial port 1 × DB-15 VGA port

Power Supply

Built-in power module with 1+1 redundant configuration

Operating System

OpenEuler Linux